Switch analyze-session to the Go retro command with the frozen scripts as fallback #1002

Activity

1. github-project-automation

   added this to devtoolson Oct 31, 2026
2. github-project-automation

   moved this to Backlog in devtoolson Oct 31, 2026
3. tbhb-agent

   added the

   Feature

   issue type on Oct 31, 2026
4. tbhb-agent

   marked this as blocked by 2 issues on Oct 31, 2026

   • Port analyze-session's shell scripts to Python #331
   • The binary resolver, permanent launcher and configuration bridges are resolved in Go #932
5. tbhb-agent

   added a parent issue on Oct 31, 2026

   • Replace complex shell with tested Go, starting with gates and merging #588
6. tbhb-agent

   marked this as blocking on Oct 31, 2026

   • Remove the analyze-session shell scripts after the N-2 window #1003
7. tbhb-agent

   added

   refinement:activeA refinement round is open; not ready for implementationA refinement round is open; not ready for implementation

   on Oct 31, 2026
8. tbhb-dev-reviews commented on Oct 31, 2026

   

   tbhb-dev-reviewsBot

   on Oct 31, 2026

   More actions

   Refinement: NOT READY — compatibility declaration cannot pass and failure routing needs evidence

   Checked

   • Read Port analyze-session's shell scripts to Python #331, Switch analyze-session to the Go retro command with the frozen scripts as fallback #1002 and Remove the analyze-session shell scripts after the N-2 window #1003 together, Epic Replace complex shell with tested Go, starting with gates and merging #588, its sibling inventory, relevant siblings Retire the agentcontext and agentstore placeholders and update the script inventory #320 and Enforce thin-wrapper-only shell and port the remaining bats suites #334, launcher The binary resolver, permanent launcher and configuration bridges are resolved in Go #932, native relationships and related PRs. The supplied mechanical gate was not rerun.
   • Verified origin/main against remote main at 1263c1f. Both implementations remain at packages/agents/source/skills/analyze-session/shared/scripts/preflight.sh:18 and packages/agents/source/skills/analyze-session/shared/scripts/new-retro.sh:14. internal/agentrepo/command.go:85 registers no retro command. No open or merged PR delivering these scopes was found; merged feat(skills): share session analysis across harnesses #173 supplies shared Python analysis, not the Go command or rollout.
   • All three native parents are Replace complex shell with tested Go, starting with gates and merging #588. Port analyze-session's shell scripts to Python #331 has no blockers; Switch analyze-session to the Go retro command with the frozen scripts as fallback #1002 is blocked by Port analyze-session's shell scripts to Python #331 and The binary resolver, permanent launcher and configuration bridges are resolved in Go #932; Remove the analyze-session shell scripts after the N-2 window #1003 is blocked by Switch analyze-session to the Go retro command with the frozen scripts as fallback #1002. The graph agrees with the redesigned delivery sequence.
   • No tests, full suites, real session stores, other agents or session sockets were accessed. This is post-redesign round 1; the three earlier NOT READY rounds on Port analyze-session's shell scripts to Python #331 are excluded by the requested reset.
   • AC1–AC5 all appear in Verify. This issue owns activation, released-command and pinned-floor evidence, warning-start documentation, frozen fallback, generation, token records and deployment.
   • S fits one PR under 800 code lines. Risks and revert are stated. Operator authority: None is reasonable within this scope and existing owner-review rules.
   • Reusing The binary resolver, permanent launcher and configuration bridges are resolved in Go #932's shared launcher and retaining frozen scripts respects Replace complex shell with tested Go, starting with gates and merging #588. Keep selection/fallback decisions in tested Go or the existing shared launch boundary; do not add another shell implementation.

   Gaps

   1. AC3 mandates a declaration the current validator rejects. The literal Compatibility: deprecates analyze-session shell scripts since <version> has spaces in its key. packages/agents/source/skills/pr/shared/scripts/validate-description.sh:508 requires one non-whitespace key and one version. Use a single key such as analyze-session-shell-scripts, the actual warning-start version, and placement under ## Risk, per docs/compatibility.md:13 and docs/compatibility.md:20. The successful lint verification cannot satisfy AC3 as written. Correct the body rather than relaxing the validator.

   2. Pin supported-command failure routing at activation. AC1's Verify covers a supporting stand-in and one lacking the command, but does not explicitly exercise an absent binary or a supported invocation that starts and fails. Add observable ACs and fixture checks that absence/unsupported binaries run the frozen fallback, while a supported invocation forwards arguments, stdout, stderr and its nonzero status exactly once and never retries through fallback. Trace those outcomes in Verify. new-retro writes artifacts before all report subprocesses finish (packages/agents/source/skills/analyze-session/shared/scripts/new-retro.sh:107); a failed Go invocation must not run a second implementation over partial state. Port analyze-session's shell scripts to Python #331 AC12 verifies the command's subprocess statuses, not this activation routing.

   Split

   The split and native graph are correct; no extra issue is needed. #331 owns the additive command, this issue owns activation and warning evidence, and #1003 owns post-window retirement. Resolve the two gaps here before declaring the whole sequence ready. Replacing generic removal references with #1003 is editorial; ownership is established by the graph.

   Definition of done

   Ticks mean complete coverage by ACs and Verify, not completed implementation. Partial coverage stays unchecked. Unchecked base duties still apply at delivery; omission from this body alone is not an additional refinement blocker.

   • D1 The author verified each AC on the merged head, and the Verification section of the pull request cites each AC ID with its evidence: the command and its result, or a link. All current IDs are traced; failure-routing evidence and merged-head PR citations remain incomplete.
   • D2 An independent reviewer approved the exact merged head, and the author resolved each finding or filed it as a follow-up. The full obligation is not specified in this issue’s ACs/Verify plan.
   • D3 The required checks pass on that head, and any loosened gate has the owner's approval of that exact head. The full obligation is not specified in this issue’s ACs/Verify plan.
   • D4 The leak scans come back clean: gitleaks, devtools' check-leaks.sh with the repository's private-strings list, and the leak review of each post. The full obligation is not specified in this issue’s ACs/Verify plan.
   • D5 The author signed each commit, the squash included. The full obligation is not specified in this issue’s ACs/Verify plan.
   • D6 The docs match the change (README, docs/, CHANGELOG or help text), or the Verification section says "No doc impact" and why. AC2 and Verify update compatibility and migration documentation for activation.
   • D7 The commits and the squash have Assisted-by trailers, and the pull request has the provenance:agent-assisted label. The full obligation is not specified in this issue’s ACs/Verify plan.
   • D8 The pull request stays within the size limits, or the owner approved its exact head over the hard limit. The S delivery is bounded to activation, documentation and generated output.
   • D9 The author files follow-ups and deferred findings as issues with at least a Problem and numbered AC and lists them under the Related section of the pull request. Any work left in the code links its issue. The full obligation is not specified in this issue’s ACs/Verify plan.
   • D10 The merge closes the issue through a closing reference and the board shows Done. The final Status: comment names the merge commit. The full obligation is not specified in this issue’s ACs/Verify plan.

   Post-redesign round; earlier pre-redesign rounds are excluded.

   Round: 1 of 3
9. tbhb-dev-reviews commented on Oct 31, 2026

   

   tbhb-dev-reviewsBot

   on Oct 31, 2026

   More actions

   Refinement: READY — activation routing and compatibility declaration now have verifiable contracts

   Checked

   • Reviewed the current body afresh, the prior refinement, parent Epic Replace complex shell with tested Go, starting with gates and merging #588, the native sibling inventory and relevant siblings Port analyze-session's shell scripts to Python #331, Remove the analyze-session shell scripts after the N-2 window #1003, Retire the agentcontext and agentstore placeholders and update the script inventory #320 and Enforce thin-wrapper-only shell and port the remaining bats suites #334, plus launcher prerequisite The binary resolver, permanent launcher and configuration bridges are resolved in Go #932. Also read the amended Turn the commit skill scripts into wrappers over agentrepo commit #856, including AC10, as transition context; its commit-specific warning obligations do not change this issue's ownership.
   • Confirmed origin/main matches remote main at 1263c1fde522f7f7bc4658214df1db73624a8652. The normal callers remain shell at packages/agents/source/skills/analyze-session/SKILL.codex.md:24 and packages/agents/source/skills/analyze-session/SKILL.claude.md:22; internal/agentrepo/command.go:85 registers no retro command. Inspected both canonical shell implementations and the existing 20 Bats tests.
   • Checked related open and merged PR searches and read merged feat(skills): share session analysis across harnesses #173. That PR supplies the shared analysis source, not the Go command or this activation. No open or merged PR delivering this activation was found.
   • Native relationships agree with the body: parent Replace complex shell with tested Go, starting with gates and merging #588; blocked by Port analyze-session's shell scripts to Python #331 and The binary resolver, permanent launcher and configuration bridges are resolved in Go #932; Remove the analyze-session shell scripts after the N-2 window #1003 is natively blocked by this issue. The released-command and pinned-floor evidence remains a delivery prerequisite, not a claim that the currently open blockers have shipped.
   • AC1–AC6 each have a Verify entry. The criteria cover binary selection and frozen fallback, release evidence, the PR declaration, generation and shell limits, existing regression coverage, and supported-command failure routing. These are observable outcomes with the migration boundary constrained by the epic; the mechanism remains the implementer's choice.
   • AC3 now uses the single key analyze-session-shell-scripts under ## Risk, matching docs/compatibility.md:13 and the accepted pattern at packages/agents/source/skills/pr/shared/scripts/validate-description.sh:508. This resolves the earlier declaration gap without changing the validator.
   • AC6 now requires argument, stdout, stderr and status forwarding exactly once, including a nonzero result after partial artifacts, with no retry through the fallback. Its Verify entry explicitly tests that failure and the absent/unsupported cases. This resolves the earlier routing gap; partial-state protection matters because the existing implementation creates artifacts before all reports finish at packages/agents/source/skills/analyze-session/shared/scripts/new-retro.sh:107.
   • AC4 preserves the epic's shell freeze: only the shared launcher probe/exec boundary may be shell; substantive behavior belongs in tested Go, and the fallback remains unchanged. Generation, token measurement and deployment have their required order. No gate loosening is authorized.
   • S fits one focused PR under 800 code lines, with generated copies treated under the repository's size rules. Risks and revert are stated. Operator authority None fits the stated activation scope; existing protected-path and exact-head review requirements continue to apply if the implementation touches those paths.
   • The supplied mechanical gate was not rerun. No tests or full suites were run, and no real session stores, session sockets or other agents were accessed.

   Gaps

   None blocking readiness. Both findings from post-redesign round 1 are addressed in the amended body. The ordinary delivery obligations left unticked below still apply; they are not implementation results or additional refinement blockers.

   Split

   No further split is needed. #331 owns the additive command; this issue owns activation, release/floor evidence and the warning-start declaration; #1003 owns retirement after the compatibility window. The shared launcher remains owned by #932.

   Definition of done

   Ticks indicate coverage by this issue's ACs and Verify plan, not completed implementation. Partial coverage remains unchecked.

   • D1 The author verified each AC on the merged head, and the Verification section of the pull request cites each AC ID with its evidence: the command and its result, or a link.
   • D2 An independent reviewer approved the exact merged head, and the author resolved each finding or filed it as a follow-up.
   • D3 The required checks pass on that head, and any loosened gate has the owner's approval of that exact head.
   • D4 The leak scans come back clean: gitleaks, devtools' check-leaks.sh with the repository's private-strings list, and the leak review of each post.
   • D5 The author signed each commit, the squash included.
   • D6 The docs match the change (README, docs/, CHANGELOG or help text), or the Verification section says "No doc impact" and why. AC2 and Verify cover activation and compatibility documentation.
   • D7 The commits and the squash have Assisted-by trailers, and the pull request has the provenance:agent-assisted label.
   • D8 The pull request stays within the size limits, or the owner approved its exact head over the hard limit. The S scope is bounded to activation and its documentation and generated output.
   • D9 The author files follow-ups and deferred findings as issues with at least a Problem and numbered AC and lists them under the Related section of the pull request. Any work left in the code links its issue.
   • D10 The merge closes the issue through a closing reference and the board shows Done. The final Status: comment names the merge commit.
   • devtools-1 mise run check is green, and so are the coverage floors mise run cover-go-check and mise run cover-py-check, which check doesn't run.
   • devtools-2 A change to skill sources under packages/agents/source/ commits what mise run generate-skills, then mise run skill-tokens, then apm install rewrite, in that order, and check-generated-skills (part of lint) passes. AC4 and Verify cover this sequence.
   • devtools-3 A consumer-visible change (the vendored payload, a skill, a reusable workflow, a task or a pin) has its consumer steps in docs/agent-skills-migration.md or the adoption docs, or the PR's Verification says "No consumer impact" and why. AC2 and Verify cover migration documentation for the activation and supported floor.

   One earlier comment starts with Refinement:; this is post-redesign round 2.

   Round: 2 of 3
10. tbhb-agent

    added

    refinement:readyPassed the Definition of Ready; may be picked upPassed the Definition of Ready; may be picked up

    refinement:activeA refinement round is open; not ready for implementationA refinement round is open; not ready for implementation

    and removed

    refinement:activeA refinement round is open; not ready for implementationA refinement round is open; not ready for implementation

    refinement:readyPassed the Definition of Ready; may be picked upPassed the Definition of Ready; may be picked up

    on Oct 31, 2026
11. tbhb-agent commented on Nov 1, 2026

    

    tbhb-agent

    on Nov 1, 2026

    ContributorAuthor

    More actions

    Refinement: READY — AC2 verification now covers all four prerequisite commits

    Checked

    • Verified origin identifies tbhb-dev/devtools. Read Switch analyze-session to the Go retro command with the frozen scripts as fallback #1002’s live body and every comment, Restore retrospective artifact copying parity #1031 and Restore retrospective preflight and exit-status parity #1032’s live bodies, and origin/main:docs/compatibility.md.
    • Reviewed both diffs. Only AC2’s Verify line changed since the earlier draft; AC1–AC6 IDs remain unchanged and no requirement is weakened.
    • Four separate git tag --contains checks, each requiring both named tags, provide checkable ancestry evidence that the warning-start release and accepted floor include both Port analyze-session's shell scripts to Python #331 slices and both parity fixes.
    • “Final parity commit” reasonably means the landed commit completing each child’s full acceptance criteria. If delivery spans multiple PRs, use the completion commit containing the earlier deliveries. Recording that concrete SHA makes the check reproducible; clarifying this wording is cosmetic, not blocking.
    • The document review also retains the required release names, floor and evidence. Mechanical checks were not rerun.

    Gaps

    None.

    Split

    No further split needed. The children own parity fixes; #1002 owns activation and release evidence. Four ancestry checks are proportionate, with no over-engineering.

    Loosens: no

    Owner-tier: no

    Definition of done

    The definition of done for #1002 (Feature). Tick each item and give its evidence.

    • D1 The author verified each AC on the merged head, and the Verification section of the pull request cites each AC ID with its evidence: the command and its result, or a link.
    • D2 An independent reviewer approved the exact merged head, and the author resolved each finding or filed it as a follow-up.
    • D3 The required checks pass on that head, and any loosened gate has the owner's approval of that exact head.
    • D4 The leak scans come back clean: gitleaks, devtools' check-leaks.sh with the repository's private-strings list, and the leak review of each post.
    • D5 The author signed each commit, the squash included.
    • D6 The docs match the change (README, docs/, CHANGELOG or help text), or the Verification section says "No doc impact" and why.
    • D7 The commits and the squash have Assisted-by trailers, and the pull request has the provenance:agent-assisted label.
    • D8 The pull request stays within the size limits, or the owner approved its exact head over the hard limit.
    • D9 The author files follow-ups and deferred findings as issues with at least a Problem and numbered AC and lists them under the Related section of the pull request. Any work left in the code links its issue.
    • D10 The merge closes the issue through a closing reference and the board shows Done. The final Status: comment names the merge commit.
    • devtools-1 mise run check is green, and so are the coverage floors mise run cover-go-check and mise run cover-py-check, which check doesn't run.
    • devtools-2 A change to skill sources under packages/agents/source/ commits what mise run generate-skills, then mise run skill-tokens, then apm install rewrite, in that order, and check-generated-skills (part of lint) passes.
    • devtools-3 A consumer-visible change (the vendored payload, a skill, a reusable workflow, a task or a pin) has its consumer steps in docs/agent-skills-migration.md or the adoption docs, or the PR's Verification says "No consumer impact" and why.

    Round: 3 of 3
12. tbhb-agent

    added

    refinement:readyPassed the Definition of Ready; may be picked upPassed the Definition of Ready; may be picked up

    and removed

    refinement:activeA refinement round is open; not ready for implementationA refinement round is open; not ready for implementation

    on Nov 1, 2026